What Is LED Lighting Made Of? Nichia 519A LED vs. Generic Chips for an Elegant Chandelier

Before you spec another elegant chandelier, let me ask the uncomfortable question: what is actually inside it?

I'm the quality compliance manager for a small lighting company. I review every LED light engine that goes into production—roughly 75 SKUs a year. Some of our fixtures end up on Perigold. Some end up in hotel lobbies. Last year I rejected about 8% of first deliveries. In almost every case, the problem wasn't an LED that failed. It was an LED that didn't match the sample I approved.

This is not a brand loyalty post. I'm comparing Option A, Nichia 519A LED boards, with Option B, generic 2835 LED boards. After four years of reviewing these components side by side, I have a clear opinion, but it comes with conditions.

What is LED lighting made of? The short version

The basic answer is an LED package, a circuit board, a driver, optics, and a heat path. The part people usually mean by 'the LED' is the package—the small yellow square on the board. Inside it, a blue chip is coated with phosphor. Blue light excites the phosphor, and the mix produces white light. That's the core trick behind nearly every modern white LED.

Once you know that, you can see why two fixtures from the same factory can look completely different. Change the phosphor thickness or the bin of the blue chip, and the white point shifts. That is exactly where the 519A and generic 2835 part ways.

The two contenders

Nichia 519A is a high-CRI mid-power package family. It's not one single SKU; you can get it in different CCTs and flux bins. If you follow Nichia LED news, you've probably seen the 519A described as a favorite for decorative lighting and retrofits. That reputation didn't come from marketing. It came from fixture designers noticing that the color, not just the CRI number, held up in real products.

Generic 2835 is the workhorse of budget LED boards. The package is 2.8 by 3.5 millimeters. You can source it from a dozen vendors, and the datasheet will often say CRI greater than 90. The question is what lands on your bench when you open the box. A datasheet is a target, not a guarantee.

Dimension 1: Color quality, not just CRI

In our lab, Nichia 519A boards measure Ra 93-95, depending on the CCT bin. A decent generic 2835 board can measure Ra 90-92. On paper, three points is not a huge difference. But CRI is an average, and averages hide weak spots.

The weak spot is red rendering, R9. I've measured 519A boards with R9 above 80. I've also measured 'CRI 90' generic boards that dropped to R9 below 20. Red rendering is what makes brass look warm instead of flat. It's what makes skin look alive under candlelight. For an elegant chandelier with visible lamps, that difference is more important than the average score.

Here's the experience override: everything I'd read about color rendering said 'pay attention to CRI.' In practice, R9 and tint consistency matter far more. I've learned to ask for TM-30 data as well. According to ANSI/IES TM-30-20, you get Rf and Rg, two numbers that describe how colors shift and how saturated they appear. The 519A tends to produce balanced Rf and Rg values. Cheap 2835 boards sometimes show a greenish Rg even when Rf looks acceptable. That's the kind of thing you'd miss if you only looked at Ra.

But here's the counterintuitive part. If the chandelier has a heavy diffuser—frosted glass, silk shade, or thick crystal that scatters light—the color rendering advantage shrinks. We ran a blind test in 2023 with a fabric shade. Our own engineers couldn't reliably tell the 519A board from a good generic board. So don't spend for color quality if the light will never touch an exposed metal finish or a human face.

Dimension 2: Tint consistency across a production run

This is where the 519A earns its keep.

LED chips are sorted into bins by brightness and chromaticity. Tight bins cost more. Standard practice is to specify a 5-step MacAdam ellipse. A premium specification is a 3-step ellipse. Nichia 519A is commonly available in 3-step bins for CCT. Generic 2835 boards, in my experience, vary unless you explicitly pay for tight binning.

In Q1 2024, we ordered 800 generic boards for a hotel chandelier project. The CRI was fine. The luminous flux was fine. But when we laid the boards side by side under a stable light source, about 12% of them looked greenish-yellow. They fell outside the 5-step MacAdam ellipse from our master sample. The vendor said it was 'within industry standard.' We rejected the whole batch. So glad we did—the replacement order was clean, but it delayed production by 11 days.

This is not a magic claim about Nichia. It is a statistical process control argument. If you order 519A from an authorized distributor, you know the bins are tight. With generic 2835, you are trusting a screen-print on the reel. For any chandelier with exposed LEDs, the 519A is the safer default.

Dimension 3: Efficiency and thermal behavior

If you compare max lumens per watt, a generic 2835 can beat the 519A. But a chandelier doesn't need maximum lumens. It needs even, warm-looking light without cooking the phosphor.

The 519A isn't the most efficient LED you can buy, but it behaves predictably at the low currents we use in decorative fixtures—usually 150 to 200 milliamps per package. In our heat tests, 519A boards ran at a lower case temperature than generic boards with the same light output. The generic boards used cheaper PCB material and smaller trace areas, which made the package run hotter. Over time, heat shifts the phosphor and changes the color point. It also reduces lumen maintenance.

That's why I ask for an LM-80 report before approving any LED. LM-80 is the IES method for measuring lumen and color maintenance. According to LM-80-15, you can compare projected lumen maintenance from sample data. Nichia publishes LM-80 data for the 519A. Many generic 2835 board suppliers either can't produce a full report or pass along vague reliability summaries. Not acceptable for a contract lighting job.

Dimension 4: Cost per fixture versus cost of rework

I'll be honest: the 519A is not a cost-saving component. On a recent job, swapping from a generic 2835 board to a 519A board raised the BOM cost by about $1.40 per fixture. That's around 30% more for the light engine. On a 500-fixture order, that's $700. Not trivial.

But here's the other side. The generic board job required an extra QC sorting step, a rejected batch, and a schedule delay. The delay alone cost more than the 519A upgrade would have. I've come to believe that total cost, not piece price, is the only honest comparison. After five years of reviewing modules, I've stopped asking 'which is cheaper?' and started asking 'which one will get us to the install date with no surprises?'

Take this with a grain of salt: those exact numbers depend on your volume, your CCT, and whether you have a decent relationship with an authorized distributor. But the pattern holds. Prevention is cheaper than rework.

The driver is the third contender

I didn't plan to add this, but I'd be lying if I said the LED package was the only decision. The driver controls flicker and dimming. A 519A won't fix a bad driver, and no LED package can guarantee compatibility with every dimmer. I've had to explain that to more than one client.

Five minutes of checking driver specs before ordering saves five days of field troubleshooting. That's not a slogan; I created a 12-point checklist after my third dimmer mismatch, and it has saved us an estimated $8,000 in potential travel and repair costs. The checklist is boring. It works.

So which should you choose for an elegant chandelier?

If a customer is browsing for a 'perigold chandelier' or an 'elegant chandelier,' they won't ask about the LED package. They'll ask about the look. But your reputation depends on what happens after the look captures them. A chandelier that makes the marble floor look slightly green is not elegant. It's a callback.

Use Nichia 519A when:

• The LEDs are visible or partially exposed.
• You're combining warm metals like brass or copper.
• You need smooth dimming below 20 percent.
• You're supplying multiple fixtures to the same space and they must match.

Use generic 2835 when:

• The light is fully diffused.
• The warranty horizon is short.
• You've verified the bin with an actual spectrometer, not just the box label.
• Budget is the deciding factor and you're willing to accept wider tint variation.

I know that's not a neat 'always choose premium' answer. I'd argue that the 519A is the safer default for an elegant chandelier, but 'safer' is not 'necessary.' Spec it for the reason, not the logo.

And next time someone asks what is LED lighting made of, tell them it's made of decisions. The emitter is just the first one.